Computing systems are currently in wide use. Some computing systems host services that respond to service requests from other computing systems (such as client computing systems).
In some examples, in order to service a request, the service computing system performs workflows by executing operations performed on the service computing system, itself (referred to as executing on-process code). These can often be performed relatively quickly. In some cases, however, in order to service a request, the service computing system needs to make application programming interface (API) calls to other, external systems (or API endpoints). The API calls to external endpoints can, sometimes, be orders of magnitude slower than executing logic in the on-process code. Therefore, the latency involved with making the API calls can make up a vast majority of the latency in fulfilling the service request. Where the API calls are made sequentially (based upon the order in which the on-service workflows are executed, or for other reasons), then the latency associated with each API call is additive in responding to the service request.
The discussion above is merely provided for general background information and is not intended to be used as an aid in determining the scope of the claimed subject matter.